09/19/23: For the latest on this subject see: FIRST NORMAL FORM - A DEFINITIVE GUIDE

Note: This is a 11/23/17 revision of Part 1 of a three-part series that replaced all of my previous posts on the subject (pages of which redirect here), in order to further tighten integration with the McGoveran formalization and interpretation [1] of Codd's true RDM.

On the one hand:

"... there is no generally accepted definition of 1NF ... the word that you see most often is 'atomic'. It is common to say that a relation is in 1NF if all its attributes [sic] are atomic ... Does 1NF equate to “atomic attribute [values]”? ... what [do] people have in mind [when they claim] atomicity? ... the [meaning] behind definitions is that you should rarely need to extract information from a value of an attribute ... But that explains why one cannot decide, depending on theory only, whether a relation is in 1NF ... it is a habitual use of data that makes attributes atomic, not theory. No wonder, there is so much mess in theory about what 1NF should be."

Karen Lopez: To bring others up to speed, Fabian is using an academic taxonomy for data modeling terms. It's valid. It's popular in research. It's not used in any major data modeling tool, nor in many practitioner resources. I have been using the industry vernaculars in my posts and here. Part of the debate that Fabian is having is because he does not tolerate the industry terms, so he chooses to attack others who use the mainstream terms. He believes they are "wrong" instead of "alternative". So that's part of the pain we have in debating his positions. I'm bilingual, but I choose to use just one set in my writings.

Fabian Pascal: Karen, bollocks. See, I can use non-academic terms too.

NULL values can be very useful, especially on indexes, as an indication of "index is not set" or "no index here", or "default inherited index applies".

I use Null values extensively (in huge database systems) with not only no problems whatsoever, but measurably signifcant advantages. People who try to tell you that "Nulls are the work of the devil", or :the sky will fall down if you allow nulls", or some such unsubstantiated childish delusion are exclusively ignorant of the correct ways to handle them. (Or too lazy/ineducable to learn their correct implementation and/or benefits.)

Fact, just plain indisputable fact. --bytes.com

I don't know that I would say that the RM is dead per se, or even holding us back. But I have to ask... why the relational model? ...A Hierarchical Model would work. In fact if we look at these NoSQL databases ... Hierarchical Models work better than relational models. The point is that some of the factors which cause one to think in terms of relationships have changed. Disk is cheap. There are definitely problems where having a strongly typed and structured model make sense. Then there are problems where not having a rigid model make sense. Structure on read seems to make sense these days. PS... Relational models don't scale. Especially on MPP.

--LinkedIn.com